B Curious
Start free
How Films Speak

The grammar under every scene · Chapter 1 of 4

One Face, Three Shots

On a face that never changed, and the audience that saw it change anyway.

4 min read

Some time around the end of the First World War, in a Moscow short of almost everything including film stock, a young director named Lev Kuleshov ran an experiment that has been retold so often it has worn smooth.

He took a close-up of Ivan Mozzhukhin, a popular actor of the pre-revolutionary cinema, looking at nothing in particular. He cut it against other shots. In one version the face followed a bowl of soup. In another, a child lying in a coffin. In a third, a woman reclining on a divan. Then he showed the results to an audience, and the audience, so the story goes, was enchanted by the actor. They praised his hunger, his heavy grief, his barely concealed desire. The face, of course, was the same face every time.

That is the version in the textbooks, and it is worth being honest about how much of it is told rather than documented. The film itself is lost. Our accounts come from Kuleshov, writing and speaking years later, and from his student Vsevolod Pudovkin, and they do not agree on the details: in Pudovkin's telling the images include a dead woman in a coffin and a little girl playing with a toy bear. Nobody kept a record of who was in the audience or what they said. Psychologists who have tried to repeat the experiment under controlled conditions have got mixed results, some finding no clear effect, others finding that context does nudge how viewers read a neutral face, though rather less dramatically than the legend suggests.

And yet the idea outlived the evidence, because every working editor knew it was true before anybody tested it. A glance recorded for one moment can be dropped into another and mean something new. A line reading that seemed flat on the day can be rescued by what comes before it. Meaning in film does not sit inside individual shots like a stone inside a fruit. It is produced in the joins.

The Soviet filmmakers of the 1920s made a doctrine of this, and then argued about it. Pudovkin tended to describe shots as bricks, laid one after another to build a meaning. Sergei Eisenstein, who had come from the theatre, thought that was far too polite. In 1923 he wrote an essay on what he called the montage of attractions, arguing for shocks aimed directly at the audience's nerves, and he came to believe that shots should collide: two images in conflict should produce a third idea that existed in neither.

Battleship Potemkin, released in 1925, put the theory on screen. Its most famous sequence shows tsarist soldiers marching down the Odessa Steps in rhythm, firing into a crowd. A mother is shot. A pram, released from her hands, bumps down the stone. Faces, boots, a pair of shattered spectacles. It is one of the most imitated passages in film history, and it did not happen. There was real violence in Odessa in 1905, and many people died, but not this massacre on these steps. The scene is built from cuts, and it was convincing enough that it is still sometimes remembered as history.

What does any of this mean for the person in front of the camera? Something humbling, and something useful.

The humbling part is that a screen performance is never wholly the actor's. It is shared with the editor, who can lengthen a pause, borrow a reaction, or cut away at the instant before a gesture overreaches. Actors who have watched themselves in a finished film often describe the strange experience of seeing a moment they do not remember playing, assembled from takes they barely registered.

The useful part is the instruction every film actor hears sooner or later, which is to do less. It sounds like a criticism of feeling, and it is not. It is a piece of arithmetic. If the cut to the coffin already tells the audience that this is grief, then a face that also strains to tell them is saying the same thing twice, and the doubling reads as false. The camera is close, the edit is eloquent, and the face only needs to be open: thinking, attending, alive. The audience will finish the sentence.

Kuleshov's lost reel proves less than it is usually credited with. The principle it stands for proves itself in almost every film you will ever watch.

Cut from a man smiling to a mother and baby and he is a kindly gentleman; cut to a girl in a bikini and he is a dirty old man.
Hitchcock demonstrating the Kuleshov effect on television in 1964, paraphrased

Where you'll meet this

Every reaction shot in every film you watch is borrowing a meaning from the shot beside it, and editors routinely use a glance recorded for one moment to play another.